{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/large-scale-language-modeling-converging-on","title":"Large Scale Language Modeling: Converging on 40GB of Text in Four Hours","arxiv_id":"1808.01371","date":"2018-08-03","proceeding":null,"authors":["Raul Puri","Robert Kirby","Nikolai Yakovenko","Bryan Catanzaro"],"abstract":"Recent work has shown how to train Convolutional Neural Networks (CNNs)\nrapidly on large image datasets, then transfer the knowledge gained from these\nmodels to a variety of tasks. Following [Radford 2017], in this work, we\ndemonstrate similar scalability and transfer for Recurrent Neural Networks\n(RNNs) for Natural Language tasks. By utilizing mixed precision arithmetic and\na 32k batch size distributed across 128 NVIDIA Tesla V100 GPUs, we are able to\ntrain a character-level 4096-dimension multiplicative LSTM (mLSTM) for\nunsupervised text reconstruction over 3 epochs of the 40 GB Amazon Reviews\ndataset in four hours. This runtime compares favorably with previous work\ntaking one month to train the same size and configuration for one epoch over\nthe same dataset. Converging large batch RNN models can be challenging. Recent\nwork has suggested scaling the learning rate as a function of batch size, but\nwe find that simply scaling the learning rate as a function of batch size leads\neither to significantly worse convergence or immediate divergence for this\nproblem. We provide a learning rate schedule that allows our model to converge\nwith a 32k batch size. Since our model converges over the Amazon Reviews\ndataset in hours, and our compute requirement of 128 Tesla V100 GPUs, while\nsubstantial, is commercially available, this work opens up large scale\nunsupervised NLP training to most commercial applications and deep learning\nresearchers. A model can be trained over most public or private text datasets\novernight.","url_abs":"http://arxiv.org/abs/1808.01371v2","url_pdf":"http://arxiv.org/pdf/1808.01371v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"large-scale-language-modeling-converging-on","repo_url":"https://github.com/NVIDIA/sentiment-discovery","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"}],"methods":[{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1808.01371","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1808.01371"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/NVIDIA/sentiment-discovery","reach":null}],"summary":{"ran_fixture":2},"by_repo_kind":{"official":{"samples":2,"ran":2,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":2,"samples":[{"code_sha256_prefix":"a1e2b64ec2a9da1d","entry":"get_valid_outs","repo":"NVIDIA/sentiment-discovery","repo_kind":"official","path":"model/model.py","file_url":"https://github.com/NVIDIA/sentiment-discovery/blob/HEAD/model/model.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"a1e2b64ec2a9da1d"}},{"code_sha256_prefix":"05751773a9708601","entry":"selector_circuit","repo":"NVIDIA/sentiment-discovery","repo_kind":"official","path":"model/model.py","file_url":"https://github.com/NVIDIA/sentiment-discovery/blob/HEAD/model/model.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"NOASSERTION","inline_ok":false,"mcp_get_code":{"code_sha256":"05751773a9708601"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}